For me, tramadol lands squarely in the gap between oral analgesics I might use for "no pain" – ibuprofen, acetaminophen – and analgesics I might use for "real pain" – hydrocodone derivatives. The literature describing the analgesic properties of tramadol is bizarre, with multiple comparisons with placebo, nerve blocks, adjunctive epidural anesthesia, etc., and very few head-to-head comparisons to the sorts of medications we routinely use. When there are comparative efficacy reports, they typically conclude that tramadol is effective...just as effective as the NSAIDs its being compared with.
The theory I've heard people use when considering use of tramadol is that it has a better safety profile than hydrocodone and is less dependence-forming. These claims may be true, but I do not believe they are true to the extent that it is clinically relevant. Tramadol still generates an opioid withdrawal syndrome and, as this article describes, overdose/abuse still results in apnea with need for ventilatory support.
Additionally, tramadol is a GABA antagonist, lowering seizure threshold. Of the 525 patients overdosing primarily on tramadol retrospectively identified at this Iranian hospital, 19 experienced apnea and 242 (46.1%) experienced seizures. This is retrospective and co-ingestants cannot be fully ruled out, but the propensity for seizure is far more surprising than the incidence of apnea.
Tramadol has a role in pain control prescribing, but, in my practice, that role is tiny.